Skip to content

Disaster tweet messages classifier into 36 categories.

Notifications You must be signed in to change notification settings

leejaeka/DisasterResponse

Repository files navigation

DisasterResponse

Disaster messages classifier webapp

Overview

Project for Data Engineering section in Udacity Data Science Nanodegree program. Dataset was provided by FigureEight which are tweets and texts that were sent during real world disasters and are labeled into 36 categories. The core idea of this project is to automate detecting unpredictable disasters so people could find their aid appropriately and quickly.

main

Examples

ex1 will output ex2 Example 2 ex2 If you input something unrelated like
ex2 This correctly output none of the categories ex2

To RUN THE APP head into 'disaster-response' folder -> 'disaster_response_pipeline_project' folder and follow instructions on README.md.
Note that .pkl file is missing due to its size (2GB). You can generate the model instead by following the instruction above.

Libraries used: pandas, plotly, sqlalchemy, sklearn, pickle, json, nltk, joblib, flask, numpy, re

Releases

No releases published

Packages

No packages published

Languages